Methods for improving uplink (UL) coverage, throughput and transmission latency are being investigated in Release 6 (R6) of the 3rd generation partnership project (3GPP). In order to successfully implement these methods, scheduling and assigning of UL physical resources have been moved from the RNC to the Node-B such that the Node-B can make decisions and manage UL radio resources on a short-term basis more efficiently than the RNC, even if the RNC retains overall control over the Node-B.
One or more independent UL transmissions are processed on an enhanced dedicated channel (E-DCH) between a WTRU and a universal mobile telecommunication system (UMTS) terrestrial radio access network (UTRAN) within a common time interval. One example of this would be a medium access control (MAC) layer hybrid-automatic repeat request (H-ARQ) or a simple MAC layer ARQ operation where each individual transmission may require a different number of retransmissions to be successfully received by the UTRAN. This operation may result in a loss of transmission sequence at the MAC layer.
In-sequence data delivery is required by the radio link control (RLC) layer for proper operation. A re-ordering function is needed to organize the received data blocks according to the sequence generated by the WTRU RLC entity. In the prior art, a data-forwarding timer, (called a T1 timer in high speed downlink packet access (HSDPA)), has been utilized for delivery of non-consecutive data blocks to higher layers. The timer is initialized when data blocks which have been correctly received cannot be delivered to the higher layers due to non-sequential reception. When the timer expires, all data blocks in the re-ordering buffer up to the next non-sequentially received data block are forwarded to higher layers.
It is difficult to properly set duration for the data forwarding timer since, even in normal operation, cell congestion occasionally generates relatively long periods between transmissions. In addition, since out-of-sequence delivery to RLC acknowledged mode (AM) may result in unnecessary retransmissions, the worst case delay must be considered in setting the data forwarding timer for proper operation. Since transmission errors are unavoidable, this timer mechanism results in increased and potentially unacceptable latency in performing RLC retransmission procedures.
In order to maintain quality of service (QoS), transmissions should not be unnecessarily delayed in providing sequential delivery. The timer mechanism is not efficient to provide proper QoS. If the duration of the timer is too short, out of sequence transmissions may be incorrectly perceived when only a particular transmission has been delayed, resulting in unnecessary requests for RLC retransmissions or potential release of the channel. If the duration of the timer is too long, detection and recovery of failed transmissions is delayed, resulting in greater difficulty in achieving low transmission latency QoS requirements. Determining the duration of the out of sequence forwarding timer is further complicated by relatively large variances in retransmission scheduling.